Postgres Pro
Downloads
Home   >   mailing lists

Re: Move PinBuffer and UnpinBuffer to atomics - Mailing list pgsql-hackers

From Alexander Korotkov
Subject Re: Move PinBuffer and UnpinBuffer to atomics
Date
Msg-id CAPpHfdu+tzn5peZuwuQ9M8mLmSC=UTM7-ARswR_cWhDnj4R5Ow@mail.gmail.com
Whole thread Raw
In response to Re: Move PinBuffer and UnpinBuffer to atomics  (Dilip Kumar <dilipbalaut@gmail.com>)
Responses Re: Move PinBuffer and UnpinBuffer to atomics
List pgsql-hackers
Tree view
On Fri, Mar 11, 2016 at 7:08 AM, Dilip Kumar <dilipbalaut@gmail.com> wrote:

On Thu, Mar 10, 2016 at 8:26 PM, Alexander Korotkov <a.korotkov@postgrespro.ru> wrote:
I don't think we can rely on median that much if we have only 3 runs.
For 3 runs we can only apply Kornfeld method which claims that confidence interval should be between lower and upper values.
Since confidence intervals for master and patched versions are overlapping we can't conclude that expected TPS numbers are different.
Dilip, could you do more runs? 10, for example. Using such statistics we would be able to conclude something.

Here is the reading for 10 runs....


Median Result
-----------------------

Client       Base          Patch
-------------------------------------------
1              19873        19739
2              38658        38276
4              68812        62075

Full Results of 10 runs...

 Base
-------------                       
 Runs  1 Client    2 Client      4 Client
-----------------------------------------------------
    1    19442        34866        49023
    2    19605        35139        51695
    3    19726        37104        53899
    4    19835        38439        55708
    5    19866        38638        67473
    6    19880        38679        70152
    7    19973        38720        70920
    8    20048        38737        71342
    9    20057        38815        71403
    10  20344        41423        77953
-----------------------------------------------------

                       
Patch
-----------                       
Runs  1 Client     2 Client      4 Client
------------------------------------------------------
    1    18881        30161        54928
    2    19415        32031        59741
    3    19564        35022        61060
    4    19627        36712        61839
    5    19670        37659        62011
    6    19808        38894        62139
    7    19857        39081        62983
    8    19910        39923        75358
    9    20169        39937        77481
    10  20181        40003        78462
------------------------------------------------------

I've drawn graphs for these measurements. The variation doesn't look random here.  TPS is going higher from measurement to measurement.  I bet you did measurements sequentially.
I think we should do more measurements until TPS stop growing and beyond to see accumulate average statistics.  Could you please do the same tests but 50 runs.

------
Alexander Korotkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company 
Attachment

pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: Recovery test failure for recovery_min_apply_delay on hamster
Next
From: David Rowley
Date:
Subject: Re: Parallel Aggregate